How alerts work

An alert is raised when a reading goes outside the limits you set for a metric. Alerts appear on the Alerts / Notifications page, and are sent to you by email and to the mobile app.

A device or gateway going offline or coming back online is a notification, not an alert.

Set the limits

Each metric can have a Min Threshold, a Max Threshold, or both.

• For a device: set them in step 4 of the Add Device wizard, or later with Edit on the Devices page. API devices have the same fields.
• For a gateway's child devices: set them under Alerts & Thresholds in step 5 of the Add Gateway wizard, or later in the gateway's edit form.

A metric with no thresholds never raises an alert.

Extra alert levels

A metric can have up to 5 thresholds in all: its Min and Max Threshold, plus up to three extra levels. Click Add alert level under a metric's thresholds, then choose:

• a severity: Critical, High or Medium,
• Above or Below, and
• the value.

An Above level must be higher than the Max Threshold, and a Below level lower than the Min Threshold. For example, with a Max Threshold of 40 °C you could add High above 45 °C and Critical above 50 °C.

A level doesn't raise an alert of its own. The alert still opens at the Min or Max Threshold, and when the reading crosses a level, that same alert is escalated to the level's severity and you're notified again. The alert's message and its Threshold field name the level it reached, for example above Critical threshold: 51 > 50.

One alert per problem

The Firmcraft IoT Dashboard raises one alert for each time a metric goes out of range, not one per reading. A voltage that stays low for an hour is one alert, not sixty.

1. Opening. A reading goes past a threshold. The alert opens when the next reading is out of range too, so one stray reading doesn't raise an alert.
2. While it lasts. Further out-of-range readings update the same alert. If the value gets worse than before, the alert is escalated and you're notified again.
3. Closing. The alert resolves itself when the value comes back inside the limit by a small margin, about 1% of the threshold. A reading hovering right on the limit doesn't open and close the alert over and over.

If the value jumps from below the minimum to above the maximum, the first alert closes and a new one opens for the other limit.

Severity

Severity depends on how far past the threshold the reading is:

Severity	Reading is past the threshold by
Critical	50% or more
High	25% to 50%
Medium	10% to 25%
Low	Less than 10%

For example, with a maximum of 40 °C, a reading of 45 °C is 12.5% over and raises a Medium alert; 62 °C is 55% over and raises a Critical one.

If you've added extra alert levels, the alert takes whichever is higher: the severity from this table, or the severity of the highest level the reading has crossed.

Offline and online

When a device or gateway stops reporting, or comes back, you get a notification, not an alert: under the bell icon, by email and as a push on your phone. It doesn't appear in the alerts list and doesn't need resolving. The IoT Dashboard learns how often each device reports, so a device isn't called offline between normal readings.

Who gets alerts

For each alert, the IoT Dashboard notifies:

• the company Admin,
• the User the device is assigned to, if any, and
• the Viewers that User invited.

Each person's own notification preferences decide whether they're emailed or sent a push notification, and whether only critical alerts reach them.

Next steps

• The Alerts page
• Assign and resolve alerts